Signal quality values, e.g. signal-to-interference ratio (SIR), are often desired to be determined in wireless telecommunication apparatuses for supporting their operation. Typically, signal quality is estimated from reference signals, e.g. pilot signals, provided in the telecommunication system. The reference signal power can thus be divided by a total interference and thermal noise, and possibly corrected using some bias compensation. However, for very low signal quality, e.g. a SIR below −3 dB, the signal quality value estimate becomes highly biased and the accuracy of the estimate is questionable. The accuracy is needed, since in a typical application, a SIR of −5 dB could be sufficient to provide a desired quality of service, while for the same application, −7 dB would disable service provision. Providing a proper bias compensation at these levels is a difficult task. Therefore, there has been discovered a need for providing improved signal quality value estimation.